May 18, 1989 - We thank our colleagues, Mark Collins, Robert DiFrancesco,. Cesar Guerra, Donald Mahan, Leslie Orgel, Donna Lee Regl,. James Stefano ...
CLIN. CHEM. 35/9, 1826-1831 (1989)
Quantitative Assays Based on the Use of Replicatable Hybridization Probes HIlda L.omell,’ Sanjay Tyagl,2 Cynthia G. PrItchard,3 Paul M. Llzardl,1and Fred Russell Kramer2’4 Amplifiable hybridization probes-molecules with a probe sequence embedded within the sequence of a replicatable RNA-wiIl promote the development of sensitive clinical assays. To demonstrate their utility, we prepared a recombinant RNA that containeda 30-nucleotide-longprobe complementary to a conserved region of the pol gene in human immunodeficiency virus type 1 (HIV-1) mRNA. Test samples were prepared, each containing a different number of HIV-1 transcripts that served as simulated HIV-1 mRNA targets.
Hybridizationswere camed out in a solutioncontainingthe chaotropicsalt, guanidinethiocyanate.Probe-target hybrids were isolated by reversibletarget capture on paramagnetic particles.The probes were then released from their targets and amplified by incubation with the RNA-directed RNA polymerase, Q/3 replicase (EC 2.7.7.48). The replicase copied the probesin an exponentialmanner:after each roundof copying,the numberof RNA moleculesdoubled.The amount of RNA synthesizedin each reaction(-50 ng) was sufficient to measure without using radioisotopes.Kinetic analysis of the reactionsdemonstratedthat the numberof HIV-1 targets originallypresent in each sample could be determined by measuringthe time it took to synthesizea particularamount of RNA (the longerthe synthesistook,the fewer the number of targetsoriginallypresent).The resultssuggestthat clinical assays involving replicatable hybridizationprobes will be simple,accurate, sensitive,and automatable. K.yphrases: recombinant RNA reversible target capture paramagnetic pailicles exponentialamplification 0$ relicase guankline thiocyanate
AddItional
.
‘
.
Infectious agents may be quite rare in asymptomatic individuals who are infectious to others. For example, an asyniptomatic individual may have as few as one in 100 000 peripheral blood mononuclear cells infected with the pathogenic retrovirus, human iminunodeficiency virus type 1 (H1V-1), yet donated blood from that person will readily infect others (1). It is thus imperative that very sensitive clinical assays be developed for detecting HJV-1, to screen donated blood and to identify asymptomatic carriers. Suitable assays would make use of a macromolecular probe having extremely high affinity for a particular component of the infectious agent and very low affinity for all the other components of the sample. The highest specificities and most stable interactions known occur when a single-stranded oligonucleotide probe hybridizes to a complementary oligonucleotide target (2). For example, ohigo1Centro de Investigacion sabre Ingemerfa Gen#{233}tica y Biotecnologla, Universidad Nacional Aut#{243}noma de Mexico, Apartado Postal 510-3,62270 Cuernavaea, Morelos, Mexico. 2Department of Molecular Genetics, Public Health Research Institute, 455 First Avenue, New York, NY 10016. 3Gene-Trak Systems, 31 New York Avenue, Framingham, MA 01701. 4Address correspondence to this author. Received May 18, 1989; accepted June 28, 1989. 1826
CLINICALCHEMISTRY, Vol. 35, No. 9, 1989
probes can seek out and bind to the integrated HIV-1 DNA, or the retroviral messenger RNA, present in a single infected cell. However, the use of oligonucleotide probes is not sufficient to assure detection. An infected cell contains only about 6000 molecules of retroviral messenger RNA (3), so the problem becomes how to detect the probes once they are bound to such a small number of targets. The classic detection strategy is to attach reporter groups to the probes, such as fluorescent organic molecules or radioactive phosphate groups. More recently, biotin groups have been incorporated into probes (4). After the probes have bound to their targets, enzymes such as peroxidase or phosphatase nucleotide
are linked
to the biotin,
then
incubated
with
a colorless
substrate, leading to the accumulation of a large number of colored product molecules for each enzyme-probe adduct (5). However, the practical limit of detection of these schemes is about 106 target molecules. Clearly, they cannot be used to detect a single cell in a sample that contains only 6000 retroviral messenger RNAs. A particularly attractive strategy for detecting rare targets is to link each probe to a replicatable reporter, which can be exponentially amplified after hybridization to reveal the presence of the probe (6). We recently described a novel version of this approach, in which a probe sequence was embedded within the sequence of a replicatable RNA (7). The resulting recombinant RNAs hybridize to their target sequences the same way as ordinary hybridization probes do and, as in a classical hybridization assay, nonhybridized probes are then washed away. The hybridized probes are then freed from their targets and released into solution. What makes these recombinant-RNA probes particularly useful is that they can then be exponentially amplified by incubation with the RNA-directed RNA polymerase, Q replicase (8). We have demonstrated that as many as 10 copies of each replicatable probe can be synthesized in a single 30-mm incubation (7). Furthermore, the extreme specificity of Q13 replicase for its own template RNA (9) assures that only the replicatable probes will be amplified. The large number of copies synthesized can easily be quantified by incorporating radioactive nucleotides or by measuring the fluorescence of an intercalating dye such as ethidium bromide. Because as little as a single molecule of RNA can initiate exponential amplification (10), this approach offers the prospect of developing sensitive diagnostic assays.
Two developments led to our current work: the discovery that oligoribonucleotides can be inserted within the sequence of a small, naturally occurring template for Q/3 rephicase, MDV-1 RNA (11), without interfering with its rephicatability (12); and the availability of a plasmid that serves as a template for the synthesis of MDV-1(+) RNA when the plasmid is incubated in vitro with bacteriophage TI RNA polymerase. We modified this plasmid by inserting a polyhinker into the MDV-1 cDNA sequence. The subsequent insertion of a probe sequence within the polylinker creates a plasmid that serves as a template for the transcription of recombinant-RNA probes (7). The site that we
chose for inserting the polylinker and probe into MDV-1 RNA was known to be on the exterior of the molecule, where the presence of the inserted sequence was less likely to interfere with replication, and where the remainder of the molecule was less likely to interfere with the hybridization of the probe sequence to its target. In our previous report (7), we demonstrated that these recombinant RNAs are bifunctional, in that they hybridize specifically to complementary target sequences and retain the ability to be exponentially amplified by Qj3 replicase. Here, we demonstrate the use of replicatable hybridization pr.obes in a model assay designed to detect very small amounts of H1V-1 mRNA. We had two main concerns in selecting the assay format: (a) because of the desirability of developing a method that can screen a large number of samples, the selected format had to be fast and simple, thus precluding the fractionation of cells or the isolation of nucleic acids; and (b) because nonhybridized probes are amplified by Q replicase along with hybridized probes, we needed an extremely efficient means of removing the nonhybridized probes. Our first problem was solved by the dual discoveries that hybridization is extremely efficient in solutions of the chaotropic salt, guanidine thiocyanate (13), and that concentrated solutions of guanidine thiocyanate will lyse cells, denature all proteins (including nucleases), liberate nucleic acids from cellular matrices, and unwind DNA molecules, permitting hybridization to occur without interference from cellular debris (3). Our second problem was solved by the development of the “reversible target capture” procedure (14). In this improved “sandwich hybridization” technique (15, 16), probe-target hybrids are bound to the surface of paramagnetic particles. After the particles are washed to remove nonhybridized probes, the hybrids are released from the particles and then bound to a new set of particles for another washing. Repeating this procedure several times dramatically reduces the number of nonhybridized probes (14). We prepared recombinant RNAs containing a 30-nudeotide-long probe sequence complementary to a conserved region of the HJV-1 pol gene. We also prepared a set of serial dilutions of a stock solution of transcripts of the region of the H1V-1 genome that contains the pol gene, to simulate different amounts of HJV-1 mRNA that might be present in clinical samples. An excess of replicatable HIV-1 probes was added to each sample. After hybridization in the presence of guanidine thiocyanate, the probe-target hybrids were isolated by reversible target capture on paramagnetic particles. The probes were then released from their targets and exponentially amplified by incubation with Qf3 replicase. Samples of each reaction were taken every minute during replication. The amount of RNA in each sample was then measured. The results demonstrate how kinetic data can be used for quantitative determination of the number of target molecules in a sample.
solution at -20 #{176}C: its activity remains unchanged after five years of storage. Single-stranded DNA fragments were prepared, by using 3-cyanoethyl phosphoramidite chemistry, on a 380A synthesizer (Applied Biosystems, Foster City, CA).
ReplicatableHIV-1 Probes Recombinant MDV-1 RNA containing an inserted H1V-1 probe sequence was synthesized by transcription from a recombinant plasmid. The plasmid was constructed by inserting a synthetic probe sequence (prepared by annealing dGATCACCGTAGCACTGGTGAAATFGCTGCCATTGA to dGATCTCAATGGCAGCAArI’TCACCAGTGCTACGGT) into the Bgl II site of a plasmid that is identical to plasmid pT7-MDV-poly (7), except that the polylinker sequence is in the opposite orientation. The nucleotide sequence in the recombinant region of the cloned plasmid was confirmed by the chain termination procedure (18). The synthesis of replicatable probes by transcription from linearized recombinant plasmids with P7 RNA polymerase is described in detail elsewhere (7). The resulting transcripts were recombinant MDV-1(+) RNAs containing a 30-nucleotide-long probe sequence that is complementary to nucleotides 4622-4651 in the pol gene of HIV-1 genomic RNA (19). Figure 1 shows the nucleotide sequence and predicted secondary structure of the transcribed RNA. MDV-hiv(+) RNA serves as an excellent template for exponential amplification by Q(3 replicase.
A A C
cc
MaterIals and Methods
C
Enzymes and Oligodeoxyribonucleotides Bacteriophage TI RNA polymerase (EC 2.7.7.6) was purchased from New England Biolabs, Beverly, MA, and calf thymus terminal deoxyribonucleotidyltransferase (EC 2.7.7.3 1) was obtained from Supertechs, Bethesda, MD. Qf replicase (EC 2.7.7.48) was isolated from bacteriophage Q/3-infected Escherichia coli Q13 by the procedure of Eoyang and August (17), with the hydroxylapatite step omitted. Qj3 replicase is stable when stored in a glycerol
C_c CCC
A C CC cC c c Cu Cu
Fig. 1. Replicatable HIV-1 hybridization probe The nucleotide sequence of MDV-hiv(+) RNA was folded intothe secondary structurespredictedto be most stable by a computer program (2. The probe sequence (bold letters) is located on the exterior of the molecule, where it Is free to hybridizeto its target, and where it Is less likely to Interfere with the sequences and structures required for replIcation (21) CLINICAL
CHEMISTRY,Vol.35, No. 9, 1989 1827
Capture Probes
Target Molecule
Single-stranded DNAs containing 3’-poly(dA) tails were synthesized for use in binding probe-target hybrids to oligo(dT) groups on the surface of paramagnetic particles. Four different oligodeoxyribonucleotides (of lengths 24, 40, 40, and 43 nucleotides) were prepared by automated synthesis. Each probe was complementary to a different region of the HIV-1 pol gene near to the target of the replicatable probe. A poly(dA) tail was added to the 3’ end of each probe by incubation with terminal deoxyribonucleotidyltransferase (22). Hybridization Simulated
HIV-1 mRNA targets were purchased from Systems, Framinghain, MA. These transcripts included a complete copy of the HIV-1 pol gene. Seven reaction tubes were prepared. Each contained simulated HIV-1 mRNA targets, MDV-hiv(+) RNA (replicatable probes), and capture probes, dissolved in 70 pL of 2.5 mol/L guanidine thiocyanate (Fluka Chemical, Hauppage, NY), and placed in a polypropylene “titertube” (Bio-Rad, Richmond, CA). Each tube contained 2 x i0 molecules of MDV-hiv(+) RNA, 1011 molecules of each capture probe, and a different number of target molecules. The number of H1V-1 transcripts in each tube was: iOn, 108, iO, 108, iOn, io, and iOu. The tubes were incubated at 37#{176}C for 18 h. Gene-Trak
ReversibleTarget Capture After the completion of hybridization, the probe-target hybrids were isolated from the reaction mixture by binding them to oligo(dT) groups on the surface of parainagnetic particles (14). These ferric oxide particles (E I
of each
sample
was bound
to a nylon
mem-
:
100.000,000
2a)
An aliquot
4
#{149} #{149} ....,.L
10,000.000
a)
1,000,000
.---c._
.,
100,000
±
#{149} I
I “,.
10,000 1,000
4
-:.!
#{149}..#{149}
.
10
12
14
16
18
20
22
24
26
28
Time (minutes) of amplification reactions initiated with replicatable probes isolated fromhybridization reactions Approximately50 ng of MDV-hiv RNA was synthesized in each amplification reaction.This amount of RNA is sufficient to have been accurately measured by the
Fig.
3. Kinetic analysis
fluorescence of an intercalating dye, such as ethidlum bromide
CLINICALCHEMISTRY,Vol.35, No.9, 1989 1829
doublings of the RNA population before there are enough RNA molecules to achieve saturation, the kinetic data can be used to calculate the number of replicatable probes that were present at the beginning of the reaction. If known standards are included among the unknown samples to be tested, then these data can be used to determine the number of target molecules originally present in each unknown sample. The results also indicate the limit of detection. The amplification reaction corresponding to the sample containing 10 targets achieved saturation at an earlier time than did the amplification reaction corresponding to the sample containing i0 targets. However, there was no significant difference in the amplification reactions corresponding to the samples containing 10 and iO targets. Accordingly, the limit of detection was about 10000 target molecules. Because electrophoretic analysis of the amplified RNA in each sample indicated that only recombinant RNA was synthesized, the limit of detection was determined by the level of persistence of nonhybridized replicatable probes. It is important to note that these were only preliminary assays, designed to demonstrate how replicatable probes might be used. Further experiments should lead to alterations in the assay format that will improve the sensitivity. DIscussion During exponential synthesis, the time it takes for the RNA population to double is a constant for a given set of reaction conditions (23). If we know how many replicatable probes were initially present in a reaction, and if we know how long that reaction was incubated, then we can predict how many doublings have occurred and how many RNA molecules have been synthesized. Conversely, if we know how long it takes for a particular number of RNA molecules to be synthesized, then we can calculate how many molecules of replicatable probe were present initially. This relationship is summarized by the following equation: N = N0
2t/d
where N0 is the initial number of RNA molecules; t is the time of incubation; d is the characteristic time it takes for the RNA population to double; and N is the number of RNA molecules present at time t. Taking the logarithm of each side of the equation and rearranging algebraically: logN0=
(_1o2)
t + logN
where log 2)/d is a constant. If we consider the situation that occurs when we determine the time it takes for each reaction to synthesize a particular number of RNA molecules, then log N will also be a constant and twill represent the time it takes for the RNA population to grow to N molecules. There will then be an inverse linear relationship between t and log N0. Therefore, if we have a reliable method for determining the time it takes for an exponentially replicating RNA population to grow to a particular (though arbitrary) number of molecules, then we can accurately determine the initial number of replicatable probes. There is a good method for determining how long it takes for a particular number of RNA molecules to be synthesized. An intercalating fluorescent dye, such as ethidium bromide, could be included in the RNA amplification reaction mix(-
1830
CLINICALCHEMISTRY,Vol. 35, No. 9, 1989
ture. Ethidium bromide becomes fluorescent when it interacts with the secondary structures present in replicatable probes. An ethidium bromide concentration of about 1 pmoll L would give a good signal, without significantly inhibiting replication A simple instrument could periodically monitor the fluorescence of the ethidium bromide in an entire set of amplification reactions. Initially, the number of RNA molecules would be too low to produce an appreciable fluorescence. However, as exponential synthesis proceeds, the fluorescence would increase. The instrument would be programmed to store the kinetic data and to use these data to determine the time for each reaction at which the fluerescence corresponds to the presence of a particular number of RNA molecules (the “endpoint”). The inclusion of standards in the hybridization reaction, each containing a known number of target molecules, would permit the establishment of a “standard curve,” in which the logarithm of the number of target molecules would be inversely proportional to the time at which the endpoint is reached (as described in the second equation). The number of target molecules in each of the unknown samples would then be determined by comparing their endpoints with those on the standard curve. This method is readily automatable; it would not require radioactive compounds; the magnitude of the fluorescent signal at the endpoint would be the same for all the reactions and would be well above the fluorescent background; the assay would be accurate; and the logarithmic nature of the standard curve would permit the determination of the number of targets in a sample over an extremely wide range of target concentrations. There is an alternative method for analyzing the data. Once the saturation point is reached in an amplification reaction, the number of RNA molecules increases linearly with time. For example, in Figure 3, by the time the reactions had been incubated for 28 mm, they had all passed the saturation point and were in the linear phase of synthesis. A comparison of the amounts of RNA present in each sample at 28 mm shows that the most RNA is present in those samples that correspond to the hybridization reactions that contained the most targets. Because these reactions were initiated with the greatest number of replicatable probes, they reached the saturation point soonest and had the longest period of time to synthesize RNA in the linear phase. We can analyze the data by using the direct linear relationship between the amount of RNA present at a particular (though arbitrary) time in the linear phase and the logarithm of the number of replicatable probes that initiate a reaction (7). If known standards are included among the unknown samples to be tested, then these data can be used to determine the number of target molecules present in each unknown sample. Because it is relatively simple to devise an assay kit to measure the amount of RNA synthesized in a reaction that is incubated for a fixed length of time, this alternative analytical approach would be an inexpensive method for detecting infectious agents in the field. (23).
The model assay we have described demonstrates that replicatable hybridization probes can be used in quantitative assays designed to detect rare targets. Further experiments will be needed to develop actual clinical assays for measuring the number of 11W-i mENA molecules in blood samples. However, it is clear that these assays will be simple, accurate, sensitive, and automatable.
We thank our colleagues, Mark Collins, Robert DiFrancesco, Cesar Guerra, Donald Mahan, Leslie Orgel, Donna Lee Regl, James Stefano, Isabel Tussie-Luna, and David Zhang, for their many intellectual and experimental contributions to the assay. This work was supported by the National Science Foundation (DMB-86-16429), the John D. and Catherine T. MacArthur Foundation, and Gene-Trak Systems. H. L. is the recipient of a predoctoral fellowship from el Consejo Nacional de Ciencia y Technologia (CONACYT). References 1. Harper ME, Marselle LM, Gallo RC, Wong-Staal F. Detection of lymphocytes expressing human T-lymphotropic virus type ifi in lymph nodes and peripheral blood from infected individuals by in situ hybridization. Proc Nati Aced Sci USA 1986;83:772-6. 2. Gillespie D, Spiegelman S. A quantitative assay for DNA-RNA hybrids with DNA immobilized on a membrane. J Mol Biol 1965;12:829-42. 3. Pelligrino MG, Lewin M, Meyer ifi WA, et al. A sensitive solution hybridization technique for detecting RNA in cells: application to HIV in blood cells. Biotechniques 1987;5:452-9. 4. Langer PR, Waldrop A.A, Ward DC. Enzymatic synthesis of biotin-labeled polynucleotides: novel nucleic acid affinity probes. Proc Natl Aced Sci USA 1981;78:6633-7. 5. Leary JJ, Brigati DJ, Ward DC. Rapid and sensitive colorimetnc method for visualizing biotin-labeled DNA probes hybridized to DNA or RNA immobilized on nitrocellulose: bio-blota. Proc NatI Aced Sci USA 1983;80:4045-9. 6. Chu BCF, Kramer FR, Orgel LE. Synthesis of an amplifiable reporter RNA for bioassays. Nucleic Acids Res 1986;14:559-603. 7. Lizardi PM, Guerra CE, Lomeli H, Tussie-Luna I, Kramer FR. Exponential amplification of recombinant-RNA hybridization probes. Biotechnology 1988;6:1197-202. 8. Haruna I, Spiegelxnan S. Autocatalytic synthesis of a viral RNA in vitro. Science 1965;150:884-6. 9. Haruna I, Spiegehnan S. Recognition of size and sequence by an RNA replicase. Proc NatI Aced Sci USA 1965;54:1189-93. 10. Levisohu R, Spiegelman S. The cloning of a self-replicating RNA molecule. Proc Natl Acad Sci USA 1968;60:866-72. 11. Kacian DL, Mills DR. Kramer FR, Spiegelman S. A replicating RNA molecule suitable for a detailed analysis of extracellular
evolution and replication. Proc Natl Acad Sci USA 1972;69:303842.
12. Miele EA, Mills DR, Kramer FR. Autocatalytic replication of a recombinant RNA. J Mol Biol 1983;171:281-95. 13. Thompson J, Gillespie D. Molecular hybridization with RNA probes in concentrated solutions of guanidine thiocyanate. Anal Biochem 1987;163:281-91. 14. Morrissey DV, Lombardo M, Eldredge JK, Kearney KR, Groody EP, Collins ML. Nucleic acid hybridization assays employing dA-tailed capture probes. I. Multiple capture methods. Anal Biochem 1989;81:345-59. 15. Ranki M, Palva A, Virtanen M, Laaksonen M, SOderlund H. Sandwich hybridization as a convenient method for detection of nucleic acids in crude samples. Gene 1983;21:77-85. 16. Syvanen A-C, Laaksonen M, S#{246}derlund H. Fast quantification of nucleic acid hybrids by affinity-based hybrid collection. Nucleic Acids Res 1986;14:5037-48. 17. Eoyang L, August JT. Q RNA polymerase from phage Q13infected E. coli. In: Cantom GL, Davis DR, eds. Procedures in nucleic acid research, Vol. 2. New York: Harper and Row, 1971:829-39. 18. Sanger F, Nicklen S, Coulson AK. DNA sequencing chain-terminating inhibitors. Proc Nati Aced Sci
with
USA
1977;74:5463-7.
19. Muesing MA, Smith DH, Cabradilla CD, Benton CV, Lasky LA, Capon DJ. Nucleic acid structure and expression of the human AIDS/lymphadenopathy retrovirus. Nature (London) 1985; 313:450-8.
20. Zuker M, Stiegler P. Optimal
computer folding of large RNA
sequences using thermodynamics and auxiliary information. Nucleic Acids Res 1981;9:133-48. 21. Nishihara T, Mills DR, Kramer FR. Localization of the Q replicase recognition site in MDV-1 RNA. J Biochem 1983 ;93:66974.
22. Nelson T, Brutlag D. Addition of homopolymers to the 3’ ends of duplex DNA with terminal transferase. Methods Enzymol 1979;68:41-50.
23. Kramer FR, Mills DR, Cole PE, Nishihara T, Spiegelman S. Evolution in vitro: sequence and phenotype of a mutant RNA resistant to ethidium bromide. J Mol Biol 1974;89:719-36. 24. Maniatis T, Jeffrey A, van deSande H. Chain length determination of small double- and single-stranded DNA molecules by polyacrylamide gel electrophoresis. Biochemistry 1975;14:378794.
CLINICALCHEMISTRY,Vol. 35, No. 9, 1989
1831
